Elad Ferber

Title: Elad Ferber: Inventor Advancing Privacy-Preserving Machine Learning

Introduction

Elad Ferber is an inventor based in Wellesley, Massachusetts, whose work addresses the protection of sensitive health information when using external machine learning models. He is associated with Synthpop Inc. and holds 1 patent.

Latest Patents

Ferber’s patent, “Systems and Methods for Anonymizing Private Data for Use in Machine Learning Models,” describes a way to use a trained external machine learning model while protecting patient information. The system identifies protected health information (PHI) in a user’s request and replaces it with synthetic PHI before sending the revised request to the external model. After receiving the model’s response, the system substitutes the original PHI for the synthetic details, producing a response that addresses the task while keeping sensitive information from being shared with the external model.

Career Highlights

Ferber’s patent reflects an approach to machine learning that combines the capabilities of external models with safeguards for private health data. His work focuses on anonymizing information during processing and restoring the relevant original details in the resulting response.

Collaborations

Ferber is associated with Synthpop Inc. His coworker, Jan Jannink, is also noted in connection with his professional network.

Conclusion

Elad Ferber’s patented work centers on enabling machine learning tasks while reducing the exposure of protected health information, highlighting privacy as an important consideration in the use of external models.